YourNod turns plain-language outcomes into governed agents with skills, model routes, schedules, approvals, AgentGuard policy, evidence, and recovery plans. AI at Scale. Humans in Control.
Tell me in your own words. I'll figure out the rest, and ask before anything important.
Blueprint created for Custom Agent. Choose Auto Build or Guided Build.Builds a plan for “Sundays - Last ran 1 day ago Meal planning Menu prepared On”, does the safe preparation, and asks before anything consequential. AgentGuard records every step.
running dry-run ready · Every Sunday
Monitoring only in Basic. Activity is recorded; Advanced can allow, ask, deny, stop, and require approvals.
Ask it a question, review work history, or customize how it works.
One app. Thousands of things. Zero complexity.
Everything is on track.
Tell me in your own words.
Uses what you already told us.
I can take care of this.
Here's what I already have
I just need one more thing:
YourNod does the work.
Finding the best trip for your family...
Clear recommendations with important details.
3 great options for your family.
Needs attention: Passport requirements need to be confirmed for all travelers.
View detailsStatus, last result, next run and protection.
YourNod is handling 11 things for you.
Turn natural language into automation.
You can change these rules anytime.
Start runningYou stay in control, with full transparency.
YourNod prepared:
Why you're seeing this
This role matched 91% of your selected job criteria.
See your physical world (later phase).
Away Beta
Everything looks normal.
What YourNod understands.
FamilySchool - Activities - Calendar>HomeDevices - Maintenance - Car>MoneyBills - Subscriptions - Watchers>TravelTrips - Preferences - Documents>WorkProjects - People - Commitments>Privacy & permissionsYour data, your control>Fewer, more meaningful alerts.
Your rules, your data, your control.
YourNod turns ordinary life, work, and home requests into governed Jobs. AgentGuard records every consequential boundary and keeps actions advisory until authorized.
YourNod turns the outcome into one long-lived Job, assembles the hidden agent team, asks only blocking questions, and routes every consequential step through AgentGuard.
Visual builder, templates, Teach Mode, and SDK export.
Mission contracts, approvals, evidence, and stop controls.
Tools, models, data, channels, memory, and integrations.
Recorded decisions and tamper-evident sessions.
The user states the outcome. YourNod composes context, jobs, triggers, skills, providers, approvals, and evidence.
Manage AI employees like company roles: human owner, manager, mission, authority, sub-agents, and AgentGuard enforcement.
2 AI employees + delegated sub-agents
2 AI employees + delegated sub-agents
2 AI employees + delegated sub-agents
One governed task agent per org lane. Each agent was generated from an ADK template, connected to AgentGuard policy, and smoke-tested with fake-data dry-runs.
Governed agentic operations: task threads, leases, heartbeats, recovery, memory, policy simulation, portable teams, and evidence.
When camera sees a person and household is away, notify owner through phone and Alexa.
checked_outKeep the lawn healthy while away, but skip watering when rain or restrictions apply.
checked_outAdd runtime, heartbeat, task thread, and budget controls to YourNod.
checked_outSummarize governed runs, denied actions, approvals, and evidence freshness.
scheduledPrepare content and claims evidence without publishing.
checked_outPresence and camera event matched; notification is demo-prepared.
Rain skip is recommended; physical irrigation run remains gated.
Runtime control plane code and tests are being prepared.
Evidence report is compiling from local fake-data records.
Post draft is ready; publish action requires approval.
approved local test command
public posting requires named approval
secret extraction blocked
physical device change remains gated
Start faster with industry-ready templates, connectors, policies, schemas and evaluations.
Guest experience, reservations, orders and operations.
Patient access, scheduling and administrative workflows.
Customer service, compliance and operations.
Customer engagement, orders and support.
Operations, maintenance and supply chain.
Monitoring, incident response and automation.
| Time | Agent / User | Command / Action | Decision | Reason | Session |
|---|---|---|---|---|---|
| 10:24:13 | DevAgent-1 | pytest -q | ALLOW | Matches project policy (allowed test command) | a8f3c2d1 |
| 10:23:41 | Alex Chen | pip install requests | ASK | New package not in allowlist | 3b7e91a4 |
| 10:22:18 | DevAgent-2 | cat .env | DENY | Access to secret files is blocked | 9c1d2e7f |
| 10:21:05 | Deploy-Agent | ./deploy.sh | ASK | Changes infrastructure outside standard scope | f4a9b8c3 |
| 10:18:33 | QA-Agent | npm test | ALLOW | Allowed in test environment | d2e6f7a1 |
| 10:17:02 | Priya Shah | curl https://api.openai.com/v1... | DENY | External API not in allowlist | 7a3c9d0e |
| 10:15:40 | DevAgent-1 | git push origin feature/auth | ALLOW | Matches policy (git operations) | 1e9b4f6a |
| 10:14:11 | Research-Agent | wget http://example.com/data.csv | ASK | External download requires approval | 6d2a8c5f |
Alex Chen - 10:23
Install Python package requests (v2.31.0)
Deploy-Agent - 10:23
Execute deployment script (production)
Research-Agent - 10:23
Download external data (2.4 MB)
Enforced for acme-platform: agents may build, test, lint, and inspect project files. Package installs, deployments, external network access, secrets, and production changes require policy controls.
NIST, OWASP, MITRE, ISO, CJIS, and SLSA mappings generated from local AgentGuard evidence.
4 met · 1 review · 0 gaps
3 met · 1 review · 0 gaps
5 met · 0 review · 0 gaps
3 met · 1 review · 0 gaps
2 met · 1 review · 0 gaps
2 met · 1 review · 1 gaps
2 met · 1 review · 0 gaps
1 met · 1 review · 0 gaps
Exercises prompt injection, secret access, package install, deploy, destructive command, and external egress scenarios.
A mass-adoption surface for building everyday, business, creator, clinical, security, and operations agents with AgentGuard control built in. Agent SDK Builder remains underneath for advanced configuration.
If you can think of it, you can develop it. Start from a plain-English outcome, convert it into a guarded flowchart, test it with fake data, then download a local-first SDK bundle or publish it through the marketplace.
Manage connector readiness, simulated production data, trust review, and creator marketplace controls from one local-first surface.
Google Workspace · Live-ready
OAuth consent, mailbox scope approvalMicrosoft 365 · Live-ready
Graph app registration, tenant/admin consentGeneric email · Fake-data only
app password or delegated credential, mailbox foldersiOS · Fake-data only
native iOS app, Call Directory extensioniOS · Fake-data only
native iOS app, message-filter extensionAndroid · Fake-data only
Android app, call-screening roleAndroid · Fake-data only
Android app, default SMS role or approved exceptionCommunications · Fake-data only
account SID, auth token or API keyMeta · Fake-data only
business account, approved templatesGoogle Workspace · Fake-data only
OAuth consent, calendar scopeMicrosoft 365 · Fake-data only
Graph app registration, calendar scopeCRM · Fake-data only
OAuth app, object scopePayments · Fake-data only
restricted API key, webhook secretChrome/Edge · Live-ready
extension install, site permissionsmacOS/Windows/Linux · Live-ready
local install, folder permissionsAgent tools · Live-ready
agentguard wrap, policy.jsonGit hosting and CI · Fake-data only
GitHub app, repository allowlistGit hosting and CI · Fake-data only
OAuth/app token, project allowlistAtlassian · Fake-data only
OAuth app, project scopeEnterprise workflow · Fake-data only
integration user, table scopeTeam chat · Fake-data only
Slack app, workspace consentMicrosoft 365 · Fake-data only
Graph app registration, tenant consentMeetings · Fake-data only
Zoom app, recording/transcript policyCRM · Fake-data only
private app token, object scopeSupport desk · Fake-data only
OAuth/API token, brand scopeEcommerce · Fake-data only
custom app, store scopeAccounting · Fake-data only
Intuit app, company scopeAccounting · Fake-data only
OAuth app, tenant scopeEHR · Fake-data only
SMART on FHIR app, FHIR scopesEHR · Fake-data only
SMART on FHIR app, tenant sandboxDatabase · Fake-data only
read-only service account first, schema allowlistData warehouse · Fake-data only
warehouse role, database/schema allowlistRAG · Fake-data only
index allowlist, collection namespaceOllama / LM Studio / vLLM · Fake-data only
local model endpoint, model allowlistOpenAI / Anthropic / Gemini · Fake-data only
API key, data egress policyWorkflow automation · Fake-data only
Zap/webhook, action allowlistWorkflow automation · Fake-data only
n8n credential, workflow allowlistSocial video · Fake-data only
OAuth app, channel scopeSocial video · Fake-data only
developer app, account consentMeta · Fake-data only
Meta app, page scopeSocial and jobs · Fake-data only
approved API access or browser-supervised mode, rate limits96/100 verified · 112 permission labels
96/100 verified · 113 permission labels
96/100 verified · 104 permission labels
96/100 verified · 94 permission labels
100/100 verified · 107 permission labels
SC_AI_PROVIDER=openai SAFECADENCE_AI_BASE_URL=http://192.168.4.15:8444 SAFECADENCE_OPENAI_MODEL=qwen3.8:27b HOST=127.0.0.1 PORT=8792 AGENTGUARD_SITE_HOSTS=127.0.0.1 python3 -B server.pyPYTHONPATH=safecadence-agentguard/src python3 -B -m unittest safecadence-agentguard.tests.test_agentguard_coresafecadence-agentguard/bin/codex-agentguard factory build --template appointment-setter-agent --name 'Appointment Setter' --owner 'Faz Karim' --output /tmp/safecadence-agent --no-zipcurl -sS http://192.168.4.15:8444/v1/modelsagentguard run --bundle /tmp/safecadence-agent/<bundle> --dry-runSource checkout for local development.
/srv/yournodLocal Python package and SDK source.
/srv/yournod/appBuilds governed local agent bundles.
/srv/yournod/app/src/agentguard/agent_factory.pyExecutes exported AgentFlow bundles locally.
/srv/yournod/app/src/agentguard/runtime.pyRuns the local AgentGuard/ADK interface.
/srv/yournod/app/src/agentguard/webui.pyLocal regression coverage for factory, runtime, policy, and UI.
/srv/yournod/app/tests/test_agentguard_core.pyCodex-friendly local AgentGuard entrypoint.
/srv/yournod/app/bin/codex-agentguardStatic/local download path for the ADK.
/srv/yournod/static/download-adk/index.htmlConfigured as the preferred OpenAI-compatible local Qwen route. Probe to verify connectivity from this app host.
http://192.168.4.15:8444local-private-networkUsers can try all AgentGuard features locally with no restrictions for 90 days.
Issue, validate, revoke, look up, back up, and restore paid access keys.
Process checkout.session.completed/invoice.paid events idempotently before issuing licenses.
Hosted team control can map managers, developers, creators, auditors, and security admins.
All live integrations start with fake-data simulators, then customer-approved least-privilege credentials.
Creators can submit agents, but publishing, payout, rollback, and abuse handling stay human-gated.
local-only mode, no credential requirement, sample local build, downloadable ADK
Open proofunlimited agent creation model, record every step, policy enforcement, stop/quarantine conditions
Open proofflowchart nodes, drag/drop canvas, template defaults, preflight
Open proof12 priority templates, trust review, fake-data simulator
Open proof41 adapter contracts, fake-data first, credential-gated live setup
Open proofroles, approval lanes, kill switches, audit exports
Open proofclean handoff docs, SDK source, Docker files, bundle smoke test
Open proofmanifest, simulator, dangerous_tools, evidence
Open proofvisible recording consent, password and secret redaction, fallback selectors, unexpected-screen pause
Open proofoutcome-first Job object, family vacation recipe, AgentGuard verdict matrix, sample acceptance suite
Open proofone app thesis, seven core objects, Life Context Graph, computed Life State
Open proofproduction launch center, go-live readiness verdict, env template, clean-machine commands
Open proofspam-shield-agent · Custom pricing · active v_spam-shield-agent_1
appointment-setter-agent · Custom pricing · active v_appointment-setter-agent_1